| |
Risk of Selecting De Novo Drug-Resistance Mutations during Structured Treatment Interruptions in Patients with Chronic HIV Infection
|
| |
| |
Clinical Infectious Diseases Sept 15, 2005;41:883-890
..... The main drawback of STI is the possible emergence of drug-resistant strains because of ongoing replication in the presence of subtherapeutic drug levels..... few studies have investigated whether these mutations were selected de novo during STI or were selected from archived mutations.... During 20 (18%) of 112 STI cycles, resistance mutations were selected; 6% of the mutations were de novo (NEW) (i.e., not detected before the start of STI), and 12% were archived mutations (i.e., mutations already detected before the STI)..... The risk of presenting mutations is significantly greater in those patients receiving antiretrovirals with a long half-life than it is in patients receiving drugs with a shorter half-life or a higher genetic barrier. Therefore, STI treatment strategies should be applied with caution in patients receiving NNRTI- or lamivudine-containing regimens. Determinations of genotypic mutations in studies of STI should be performed soon after the interruption of HAART to increase the likelihood of detection.....
Authors: M. Arnedo-Valero,1 F. Garcia,2 C. Gil,1 T. Guila,1 E. Fumero,2 P. Castro,2 J. L. Blanco,2 J. M. Miro,2 T. Pumarola,1 and J. M. Gatell2
1Microbiology Laboratory and 2Infectious Diseases Unit, Institut d'Investigacions Biomediques August Pi I Sunyer, Hospital Clinic, Faculty of Medicine, University of Barcelona, Spain
For many patients the short term risks of HAART like side effects and body changes and requirement for strict adherence may outweigh the long-term benefits. Partly in response to these concerns, structured treatment interruptions (STIs) are being considered in a variety of settings.
Although the benefits of STI remain controversial and poorly defined, there is general consensus that the strategy involves an element of risk. The interruption of an effective regimen is almost invariably associated with a rapid increase in plasma HIV RNA levels and a rapid decrease in peripheral CD4+ T cell counts. Among patients responding well to HAART (i.e., those with plasma HIV RNA levels of <50 copies/mL), there is a risk that treatment interruption will select for drug-resistant strains of HIV. Theoretically, suboptimal levels of drugs soon after treatment interruption may allow viruses to replicate, thus leading to the emergence of drug-resistance mutations.
To date, several studies of the selection of drug-resistant HIV mutant strains during STI have tested whether these mutations were induced de novo (new) by STI or were archived mutations that emerged during interruptions of HAART. The study of genotypic resistance mutations in STI trials at several time points (including in samples obtained at baseline [i.e., before receipt of any treatment] and before the STI) can establish whether resistance mutations are induced by STI and whether mutations selected before the STI had any influence on the accumulation of new mutations or on the response to HAART after the STI.
ABSTRACT
Background. Structured treatment interruption (STI) may allow viral replication in the presence of decreased plasma drug levels, with risk of selection of resistance mutations.
Methods. For patients recruited for an STI study, genotypic resistance testing was performed at baseline (before receipt of any treatment), immediately before the STI, and 2 weeks after each interuption of therapy.
Results.
During 20 (18%) of 112 STI cycles (95% CI, 11%-26%), resistance mutations were selected; 6% of the mutations were de novo (NEW) (i.e., not detected before the start of STI), and 12% were archived mutations (i.e., mutations already detected before the STI).
Overall, resistance mutations during STI were selected in 9 (26%) of 35 patients; 5 (14%) of the mutations were de novo, and 4 (12%) were archived mutations.
Mutations conferring resistance to nonnucleoside reverse-transcriptase inhibitors (NNRTIs) were selected in 3 (23%) of 13 patients receiving NNRTI-based regimens (all mutations were de novo).
Mutations conferring resistance to lamivudine were selected in 9 (50%) of 18 patients receiving lamivudine-containing regimens (4 [22%] were de novo, and 5 [28%] were archived mutations). Mutations conferring resistance to nucleoside reverse-transcriptase inhibitors (NRTIs), excluding the M184V mutation, were selected in 2 (6%) of 35 recipients of NRTIs (1 [3%] of these mutations was de novo, and 1 [3%] was an archived mutation.
Finally, mutations conferring resistance to protease inhibitors were selected in none of the 22 patients receiving protease inhibitors. In most cases, de novo and archived mutations were selected during the first STI cycle, and their number did not increase during successive cycles. Plasma viral load decreased to undetectable levels in all the patients when the earlier drug regimen was reintroduced.
Conclusions. Genotypic mutations are selected during STI in a high proportion of patients (especially in patients receiving NNRTIs or lamivudine). Approximately one-half of selected mutations were archived mutations. Patients who had archived mutations did not have a higher risk of accumulating new mutations than did patients who were infected with wild-type virus before the STI.
There were no statistically significant differences in baseline characteristics, type of HAART received, number of STI cycles, or outcome between the 35 patients randomly selected to undergo genotypic studies and the 35 patients who were not selected (data not shown). Patients underwent a median of 3.2 cycles of STI, separated by periods of 2-16 weeks in which they received the original drug regimen. All of the patients (except patient 214) had good adherence, taking >95% of the prescribed dose during the study period, as measured with the remaining pill count method.
Genotypic mutations detected during STI.
At baseline (i.e., before receipt of any antiretroviral therapy), no genotypic drug-resistance mutations were detected. Overall, drug-resistance mutations were selected in 9 (26%) of 35 patients during STI. Five (14%) of the drug-resistance mutations were de novo, and 4 (12%) were archived mutations. Drug-resistance mutations were selected in 20 (18%) of 112 STI cycles (95% CI, 11%-26%); in 7 cycles (6%), the drug-resistance mutations were de novo, and in 13 (12%), they were archived mutations detected before the STI. In most cases, mutations were selected during the first STI cycle, and their number did not increase significantly during successive cycles, even for patients who had archived mutations.
All genotypic analyses were performed on samples obtained 2 weeks after interruption of therapy. In a subgroup of 8 patients, genotypic and phenotypic analyses were also performed on samples obtained at peak viral load after each interruption; none presented evidence of genotypic or phenotypic drug resistance. However, when samples obtained 2 weeks after STI were analyzed, the de novo mutation M184V was selected in strains obtained from 2 of these 8 patients (patients 4 and 211) (table 2). Despite the emergence of mutations before STI and the selection of mutations during the STI period, the viral load in plasma decreased to undetectable levels in all of the patients when the earlier drug regimen was reintroduced after each interruption of HAART.
Mutations associated with NRTIs.
All patients received NRTI-based regimens. Mutations conferring resistance to NRTIs (excluding M184V) were selected in 2 (6%) of 35 patients. In 1 (3%) of the patients (patient 241), the resistance mutation was de novo, and in 1 (3%) of the patients (patient 214), it was an archived mutation. Performing a similar analysis for the STI cycles, we found that 3 (3%) of 112 cycles were associated with the selection of mutations conferring resistance to NRTIs (excluding M184V). For 2 (2%) of the cycles, the resistance mutations were de novo, and for 1 (1%), the resistance mutation was an archived mutation.
Eighteen patients received a lamivudine-containing regimen. Mutations conferring resistance to lamivudine-mainly M184V-were selected in 9 (50%) of 18 patients; 4 (22%) of the patients had resistance mutations that were de novo, and 5 (28%) had resistance mutations that were archived mutations (figure 1C). Performing a similar analysis for STI cycles, we found the mutation M184V was selected in 18 (36%) of 50 cycles. In 6 (12%) of the cycles, the resistance mutation was de novo, and in 12 (24%), the resistance mutation was an archived mutation. In 3 (8.6%) of 35 patients (patients 248, 191, and 199), no resistance mutations were detected during STI, despite the fact that these patients presented the M184V resistance mutation before being included in this STI trial.
Mutations associated with NNRTI.
Thirteen patients received an NNRTI-containing regimen. Mutations conferring resistance to NNRTIs were selected in 3 (23%) of 13 patients; these mutations were selected de novo during STI in all 3 patients. When a similar analysis was performed for cycles of STI, we found genotypic mutations selected in 4 (27%) of 15 cycles of STI (all mutations were selected de novo).
Mutations associated with PIs.
Twenty-two patients received a PI-containing regimen. We did not find resistance to PI drugs in any patient during or before STI cycles. Eight patients showed protease polymorphisms (L63P, A71V, L10I, and M36I) during treatment interruptions that existed at baseline (table 2 and figure 1E). The relative risk of selecting resistance mutations associated with regimens containing NNRTIs versus those containing PIs was 9.82 (95% CI, 5.61-17.19; P = .0002).
DISCUSSION
The main beneficial effect of STI could potentially be to decrease long-term toxicity and adverse effects and to reduce costs. This is probably the main goal of current STI trials, because lasting induction of protective immune responses could not be achieved in various STI trials. The main drawback of STI is the possible emergence of drug-resistant strains because of ongoing replication in the presence of subtherapeutic drug levels. The present study was undertaken to analyze the risk of developing de novo drug-resistance mutations during STI and to study whether mutations selected before STI (i.e., archived mutations) have any influence on the accumulation of new mutations or on the response to the original HAART regimen when it is reinitiated after STI. Resistance mutations were selected in 9 (26%) of 35 patients who underwent STI, a proportion similar to that reported by other investigators [15]. However, few studies have investigated whether these mutations were selected de novo during STI or were selected from archived mutations [11, 15]. To properly assess whether drug-resistance mutations are induced as a result of STI, we should remember that viruses amplified during the viral load rebound that occurs after interruption of antiretroviral therapy may be ancestors with previous mutations that reemerge because of the loss of antiretroviral pressure. Metzner et al. [11] found that, at different times during STI, the M184V mutation was detected as a minor population in strains recovered from 14 of 25 subjects, and the L90M mutation was detected in strains recovered from 3 of 25 subjects. Metzner et al. [11] suggested that HIV variants carrying drug-resistance mutations may emerge during periods of increased HIV replication. In our study, selected mutations were already present in strains recovered from approximately one-half of the subjects before they were recruited for participation in the STI study. The use of more-sensitive tests for detection of mutations, including mixtures, might have more adequately distinguished between mutations classified as archived according to the definitions in the current study and those mutations that had already been selected but might have been present at levels too low to be detected by the methods employed. This point represents a caveat in regard to the interpretation of our results.
Moreover, patients with archived mutations that were selected during STI (the M184V mutation in all cases) did not have a higher risk of accumulating new mutations than did patients who had wild-type virus before STI (table 2). In most cases, mutations were selected during the first STI cycle, did not significantly increase in number during successive cycles, and were intermittently present. As Papasavvas et al. [12] reported, detection of drug resistance during STI did not predict persistence of a resistant viral population during subsequent STIs. The response to therapy was always good, with a decrease in plasma viral load to undetectable levels in all patients when the original HAART regimen was reintroduced. These findings are in agreement with those reported by other investigators [10, 12].
In an earlier pilot study of STI, we found neither genotypic nor phenotypic evidence of resistance [17]. In that study, viruses were amplified at the peak viral load after treatment interruption, and all viruses were wild type. However, in strains recovered from 2 of the 8 patients for whom samples were analyzed 2 weeks after STI, we found the M184V mutation. One of the possible reasons for the loss of this reemerging mutation at peak viral load could be the impaired fitness that it confers on the virus, as other investigators have suggested, in patients infected with multidrug-resistant HIV who are undergoing STI [22, 23]. Although the estimates of the fitness of M184V HIV mutants vary considerably, depending on laboratory methodology and the viral strain used, the replication efficiency of these viruses appears to be 3%-10% lower than that of wild-type HIV [24].
As other investigators have reported [10, 11, 15], the most frequently selected mutation during repeated treatment interruptions was M184V. Among 13 recipients of NNRTIs and 18 recipients of lamivudine, the proportions of patients who had mutations were 23% (with NNRTI mutations) and 50% (with the M184V mutation), respectively. Selected mutations were already present before recruitment for STI protocols in strains recovered from approximately one-half of the subjects receiving lamivudine-containing regimens but were present in none of the strains recovered from the patients taking NNRTIs. These data suggest that STI treatment strategies should be applied with caution in the group of patients receiving NNRTI- or lamivudine-containing regimens. The relative importance of intracellular half-life should also be taken into account when therapy with an antiretroviral drug has to be discontinued. Because mutations were sometimes selected during the first STI cycle, subjects who are taking a break from their NNRTI-based combination should consider discontinuing NNRTI therapy several days before stopping therapy with the other anti-HIV drugs.
During this study, nucleoside analogue mutations and PI-associated resistance mutations were rarely or never selected during STI; only some polymorphisms in the protease region of the pol gene were selected at baseline, before any antiretroviral therapy was received or STI was started.
In summary, our data suggest that approximately one-half of the drug resistance-associated mutations selected during STIs were mutations that had previously been identified in these individuals and that had been archived. However, selected novel mutations were also found in some individuals during the STIs, and this apparently occurred to a greater extent than might have happened among individuals who continued to receive therapy or among individuals who harbored wild-type viruses, although this conclusion has to be taken with caution, because there was not a control group. Patients who had archived mutations during STI did not have a higher risk of accumulating new mutations than did patients who had wild-type virus before STI, and mutations that were selected were intermittently present during successive cycles. Determinations of genotypic mutations in studies of STI should be performed soon after the interruption of HAART to increase the likelihood of detection. The risk of presenting mutations is significantly greater in those patients receiving antiretrovirals with a long half-life than it is in patients receiving drugs with a shorter half-life or a higher genetic barrier. Therefore, STI treatment strategies should be applied with caution in patients receiving NNRTI- or lamivudine-containing regimens.
Patients.
Thirty-five of 70 patients with chronic HIV infection who were recruited into 5 different STI pilot studies were randomly selected for an analysis of genotypic resistance mutations at several time points. Patients had received HAART for >1 year, had a plasma viral load of <20 copies/mL for >32 weeks, and had a nadir CD4+ cell count of >500 x 106 cells/L. Some of the patients had started antiretroviral therapy with 2 nucleoside reverse-transcriptase inhibitors (NRTIs), because this therapy was started before protease inhibitors (PIs) were available. All of the patients were receiving HAART (either 2 NRTIs and 1 PI or 2 NRTIs and 1 nonnucleoside reverse-transcriptase inhibitor [NNRTI]) at the beginning of the STI trial. During a 1-year period, the patients discontinued HAART 2-4 times during fixed time intervals (2-16 weeks). A total of 112 cycles of STI in 35 patients involved in STI protocols were studied. Adherence was measured with the remaining pill count method.
Plasma samples were analyzed at baseline (i.e., before start of any treatment), before the start of the STI, and 2 weeks after each interruption of therapy. Archived mutations were defined as those mutations already detected before the start of the STI. For samples obtained after initiation of HAART and before STI, resistance testing was performed on the plasma sample collected immediately before the first sample in which the HIV load was undetectable. All drug regimens were interrupted at the same time point according to the STI protocol. In a subgroup of 8 patients, genotypic and phenotypic analyses were also performed on samples obtained at peak viral load after each interruption. Plasma HIV RNA levels were determined using the Amplicor HIV Monitor Ultra Sensitive Specimen Preparation Protocol Ultra Direct Assay (Roche Molecular Systems), with a limit of quantification of 20 copies/mL. Population-based genotypic resistance testing was performed with use of the TruGene Assay (Visible Genetics).
|
|
| |
| |
|
|
|
